Late last month, Anthropic released a report arguing that GLM-5.3, a powerful model released in August by the Chinese company Z.ai, poses a major cyberattack risk because it is freely available for anyone to download and modify. The claim is the latest move in a long-running debate about whether frontier AI models should be open or locked behind the doors of frontier labs.
Much of the debate is playing out in public, and a lot of money rides on the outcome. Nvidia CEO Jensen Huang, who stands to gain if people keep buying their own hardware to run open models, enthusiastically advocates for the United States to develop more of them. Anthropic's Dario Amodei, whose company could be undermined by cheap, powerful models, advocates a more cautious approach. Beneath these commercial incentives, however, there is real substance. As David Gray Widder of the University of Texas at Austin puts it, how someone views openness is shaped by how they view AI overall.
Those concerned about AI safety warn that open models can easily be misused by bad actors. Others focus on AI's practical potential and argue that this power should not sit behind the closed doors of a few big companies, and that open models could spread it more widely. Options range from banning open models entirely to releasing them freely, and any approach involves trade-offs between safety and democratization.
Open models offer concrete benefits. Closed models such as those powering ChatGPT and Claude are accessed only through a website, app, or API. Open models can be downloaded, with their model weights publicly available, letting people run them on their own computers and, given expertise, retrain them. Closed models are currently more capable, with the top eight models on the Artificial Analysis LLM leaderboard all closed, though open models trail by only a few months. Researchers depend on open models to observe internal workings and chains of thought, combine them, compress them, and fine-tune them. Businesses have reasons to prefer them too.
Source: MIT Technology Review AI · Summarized by HeadlinesBriefing